content identification for audiovisual archives · 2019. 3. 6. · •there would be one global...
TRANSCRIPT
![Page 1: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/1.jpg)
Content Identification for Audiovisual Archives
![Page 2: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/2.jpg)
• There would be one global audiovisual archive • It would hold every work ever made • Along with every significant version • Every asset would have:
• Complete descriptive and technical metadata • Linked assets for key art, critical reviews, etc.
• Convenient search for specific works • Intelligent discovery for related works
In an Ideal World…
![Page 3: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/3.jpg)
This is not an ideal world…
but we can make it better.
![Page 4: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/4.jpg)
No Archive is an Island ● Archives do not remain static or operate in isolation
○ You obtain new audiovisual assets ■ Including new copies/different versions of works already held
○ You obtain new collateral materials
■ Key art, one sheets, scripts, critical reviews, etc.
○ You acquire new/expanded/improved metadata
○ You cooperate with others to fill gaps in local coverage
● Each of these transactions involves title matching ○ A traditionally manual (and increasingly costly) process
● There is a limit to how efficient and cost effective any process can be with manual touch points.
![Page 5: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/5.jpg)
Preserving Our Cultural Heritage
“Archives exist for the preservation and continuation of the cultural heritage … [their collected assets] require saving, gathering, preserving and/or conserving and they also need to be accessible to encourage the spread of knowledge.”
— UNESCO Audiovisual archives: A practical reader
• Accessibility requires useful and accurate Search and Discovery
• Search and Discovery require useful and accurate Metadata.
The United Nations Educational, Scientific and Cultural Organization
![Page 6: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/6.jpg)
Identification Makes It All Possible
“[M]etadata is increasingly defined as a conjoining of disparate, distributed resources. An independent numbering system is required that uniquely identifies individual [television] show and movie releases.”
— nScreenMedia nScreen nSights: TV Metadata in Transition
![Page 7: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/7.jpg)
Metadata-Based Record Matching • Metadata-based matching is easy for people to do
– Common Fields: Title, Release Date, Cast, Synopsis … – Data vary significantly by Source (and often over time)
– People are expensive and inconsistent
• Metadata matching can be automated to a point – As with EIDR automated de-duplication system
• Manual record matching cannot be avoided – Many current workflows require multiple touch-points and manual matching
with each iteration
• The goal is to match once and link all future workflows with a shared identifier
![Page 8: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/8.jpg)
Common Identification Systems • Shelf Numbers (Physical)/Directory Paths (Digital)
– Most useful for locating an asset once identified
• Classification Systems – Used by libraries, including the Library of Congress
• Proprietary Inventory Indexing – Popular with archives, including the British Film Institute
• Statistically-unique Identifiers – UUIDs, UMIDs, File Hashes (limited to digital works)
• Shared, Curated, Globally-Unique Identifiers
– EIDR
Pri
vate
Iden
tifi
ers
Glo
bal
ly-U
niq
ue
IDs
![Page 9: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/9.jpg)
Bringing It All Together with EIDR IDs
![Page 10: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/10.jpg)
Anatomy of an EIDR ID
● EIDR is a Digital Object Identifier (DOI)
○ ISO Standard 26324:2012
○ Applicable to abstract, physical, and digital assets
● The EIDR Content ID registry supports 1,208,925,819,614,630,000,000,000 unique records.
10.5240/ XXXX-XXXX-XXXX-XXXX-XXXX-C
Standard DOI prefix for EIDR Content IDs
Unique suffix for each asset
Check character
![Page 11: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/11.jpg)
Standard Theatrical Asset Types
Movie (or Short) Abstract “title” record with Derived Edits, Clips, and Manifestations plus Optional
Franchise collections
Franchise / Brand
Movie/TV /Short/etc.
Creative Version
Technical Version
Clip
Object / Track
Compilation
Abstraction Edit
Clip
Manifestation Track
![Page 12: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/12.jpg)
Standard Broadcast Asset Types
Franchise / Brand
Season / Flight
Episode / Event
Series / Program
Module / Cart /
Segment
Event / OTO / Movie
Creative Version
Technical Version
Clip
Object / Track
Airing Airing
Content ID Service ID Date/Time }
Compilation
Series
Season
Episode
TV / Short / Movie
Segment Edit
Clip
Manifestation Track
Television/Home Entertainment assets are more complex, but each is still
identified by a unique EIDR ID
![Page 13: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/13.jpg)
EIDR Best Practices ● EIDR Best Practices provide guidance for:
○ Different Work Types
○ Inheritance (Hierarchical) Relationships
○ Other Relationships & Groupings
● New Practices are developed based on member need
Scripted Fiction Documentaries Actualities
Features/Shorts/OTO TV Series/Seasons/Episodes Serials
Edits/Cuts Creative Versions Clips
Manifestations Technical Versions
Compilations Composites Collection/Franchise
Promotional Items Supplemental Items Alternate Content
![Page 14: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/14.jpg)
Case Study: British Film Institute’s
UK Filmography
![Page 15: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/15.jpg)
British Film Institute’s UK Filmography ● Some countries (e.g., France) mandate a national
filmography (usually to support taxation schemes) ● The UK does not have this mandate – so the BFI created
of a digital UK filmography ● The BFI catalog now lists every fiction feature film ever
produced in the UK ○ Based on Gifford’s British Film Catalogue plus all films since 2000
(Gifford’s last release year) ○ Pre-1930, every fiction film made in the UK ○ Post-1930, every fiction film ≥ 40 minutes (“features”) ○ Each Filmography record is matched to an EIDR ID
![Page 16: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/16.jpg)
BFI’s UK Filmography and EIDR
● BFI database records were added from various sources for various purposes over time
● Data standards and encoding practices varied historically, leading to gaps in metadata coverage and an increased risk of record duplication
● Registration with EIDR led to “a systematic deduplication and data improvement process in the BFI database”
● The EIDR-matched and cleaned database is now used for various internal and external initiatives, including lobbying for archive deposits
![Page 17: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/17.jpg)
ui.eidr.org
![Page 18: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/18.jpg)
collections-search.bfi.org.uk
![Page 19: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/19.jpg)
www.imdb.com
![Page 20: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/20.jpg)
BFI’s UK Filmography and EIDR, cont. “[T]he degree of certainty EIDR registration provides about the uniqueness of our work records is a major benefit to this data project.
Although [this project] could be achieved independently of EIDR Registration, the tools, resources and expertise provided by EIDR have catalysed it and facilitated it.”
— Stephen McConnachie, Head of Data, British Film Institute
![Page 21: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/21.jpg)
EIDR-Enabled Use Cases for the Global Archive
Community
![Page 22: Content Identification for Audiovisual Archives · 2019. 3. 6. · •There would be one global audiovisual archive • It would hold every work ever made • Along with every significant](https://reader036.vdocuments.us/reader036/viewer/2022063008/5fbce684477d391e505efe75/html5/thumbnails/22.jpg)
EIDR ID-Enabled Applications
● Any situation where you need to link materials obtained from or belonging to multiple parties
● Any workflow that currently involves manual intervention for work identification
● For example: ○ Federated Search & Discovery ○ Linking Enhanced Metadata to Off-Air Recordings ○ Bulk Digitization of Legacy Assets ○ Intellectual Property Rights Repositories